Journal of Cancer 2016, Vol. 7 671

Ivyspring International Publisher Journal of Cancer 2016; 7(6): 671-680. doi: 10.7150/jca.13684 Research Paper Sex Difference of Egfr Expression and Molecular Pathway in the Liver: Impact on Drug Design and Cancer Treatments? Lishi Wang1,2, Jianqi Xiao1,3, Weikuan Gu1, Hong Chen3 

1. Department of Orthopedic Surgery and BME-Campbell Clinic, University of Tennessee Health Science Center, Memphis, Tennessee, 38163, USA. 2. Department of Basic Medical Research, Inner Mongolia Medical University, Inner Mongolia, 010110, China. 3. Center of integrative research, The first Hospital of Qiqihaer City, 30 Gongyuan Road, Longsha District, Qiqihaer, Heilongjiang, 161005, PR China.

 Corresponding authors: Weikuan Gu, 956 Court Ave, Memphis, TN 38163, USA. Tel: 1-901-448-2259; Email: [email protected]; Hong Chen, 30 Gongyuan Road, Longsha District, Qiqihaer, Heilongjiang, 161005, PR China. Tel: 86-0452-2425981; Email: [email protected].

© Ivyspring International Publisher. Reproduction is permitted for personal, noncommercial use, provided that the article is in whole, unmodified, and properly cited. See http://ivyspring.com/terms for terms and conditions.

Received: 2015.08.28; Accepted: 2016.01.22; Published: 2016.03.20

Abstract Epidermal growth factor receptor (EGFR) has been used as the target in drug design for cancer treatment including the liver cancer. Men and women have different levels of EGFR expression during the life. The whole genome expression profiles of livers of recombinant inbred (RI) strains derived from C57BL/6J (B6) X DBA/2J (D2) were used to compare three major molecular aspects of Egfr : the relative expression levels, gene network and eQTLs that regulate the expression of Egfr between female and male mice. Our data suggest that there is a significant difference in the expression levels in the liver between female and male mice. Several important in the gene network of Egfr are differentially expressed between female and male mice. The regulatory elements for the expression levels of Egfr between female and male mice are also different. In summary, our data reveals an important sex difference in the Egfr pathways in the liver of the mice. These data may have substantial impact on drug development and dosage determinant for women and men in the clinical trials.

Key words: Cancer, Egfr, liver, Mice, Pathway, Sex.

Introduction Epidermal growth factor receptor (Egfr) is during life [5-6]. Mouse models have also suggest that involved in tissue development and in a variety of accelerates hepatocellular carcinoma progression in a cellular functions, such as cell proliferation, sex-dependent manner in a mouse model of differentiation, motility, and survival [1]. Most hepatocarcinogenesis [7-8]. These data indicated that importantly, it has a significant role in cancer gender difference may affect the cancer development, development [2]. It is highly expressed in liver and cancer treatment, and cancer survival. also regulated by estrogen [3]. Therefore, it is Several drugs for the treatment of important to understand its molecular pathways in hepatocarcinomas have been developed by targeting liver and similarities and differences between men the Egfr. For example, erlotinib, an epidermal growth and women. factor receptor and tyrosine kinase inhibitor, is a There are considerable reports on the role of Egfr targeted drug that was approved for the treatment of in the cancer development. Recently, Lanaya et al. [4] hepatocellular carcinomas [9]. Lapatinib is an reported that the presence of Egfr-positive liver inhibitor of both epidermal growth factor receptor macrophages in Hepatocellular carcinoma patients is and HER2/NEU, whereas both of which are associated with poor survival. Accompanied is the implicated in hepatocarcinogenesis [10]. While the fact that men and women have different levels of Egfr drugs based on anti Egfr are in development and in

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 672 clinical trials, its gender specificity should have been method reported previously [14]. R values were thoroughly investigated. In order to obtain compared between the association of female and male meaningful data, the study of sex-specific molecular mice. We followed the standard criteria for the strong, pathways needs samples of populations of both sexes correlation, and none correlation. When R value is of the same age, genomic background, and equal or more than 0.7 or -0.7, we regard the environment. It is extremely difficult to collect such a correlation as strong positive or negative. When R population from humans. Therefore, utilizations of value is between 0.35 and 0.69 or -0.35 and -0.69, the animal models have been as the alternative approach correlation exists but not strong. Any R values for the study of the gender difference of humans [11]. between 0 and 0.35 or 0 and -0.35 will be treated as The purpose of this work is to systematically none correlation [14]. The gene networks were investigate the sex differences of Egfr using high constructed using tools in the GeneNetwork. We quality data from the liver of a mouse population of constructed the gene network based on the Network recombinant inbred (RI) strains derived from Graph. Network graph is often used to visualize C57BL/6J (B6) X DBA/2J (D2) [12-13]. multiple sets of interactions. It consists of nodes and edge connecting nodes. In addition, the network Materials and Methods graph can represent the strength of the interaction by the thickness of the edge—that is, the higher the Data sets for analysis of number of interactions between the two nodes, the profiles thicker the edge becomes. UNC Agilent G4121A Liver Database (Jul04) was used for this analysis (http://www. Mapping transcriptomic loci (eQTL) that genenetwork.org/webqtl/main.py). The data set regulates the expression level of Egfr in female includes gene expression in livers of naive mice of and male mice both sexes from C57BL/6J, DBA/2J, B6D2F1, and 37 Transcriptome mapping with GeneNetwork was BXD strains using Agilent G4121A [12-13 ]. The data conducted to identify the chromosomal regions that sets at GeneNetwork are publically available and they affect the expression of Egfr in female and male mice. were validated using sex specific probe sets such as After we obtained the transcriptome maps, we Xist and Dby. In general, sex-balanced samples (one compared them to see whether there is a difference array from male cases, one array from female cases) between maps of female and male mice and among from RI strains derived from C57BL/6J (B6) X different tissues. We focused on the loci that have LRS DBA/2J (D2) and parental strains were used. While score equal to or higher than the significant levels. If they are collected at the similar age, performed with we did not find the loci at the significant levels, we the same microarray platform, and grown in the same then looked at the loci at the suggestive levels [14-15]. environment, most of them are from different RI For the loci showing sex difference, we strains. However, these RI strains are all derived from examined the candidate genes that regulate the Egfr the same two progenitors. Also, the gene expression expression level. We focused on the genes under the profiles of both female and male of two progenitors peak region of the eQTL, which usually is the are generated separately. Additional details regarding region with LRS scores at the significant animals, microarrays, data acquisition, processing, or suggestive levels [14]. and analyses can be found at WebQTL.org After we identified the genes in the chromosome (http://webqtl.org/dbdoc/LV_G_0106_B.html). region under the peak region of eQTL, we searched Collections of data on expression levels of Egfr the literature report to see whether any of these genes have been reported to regulate or functionally and associated genes associate with Egfr. We used the PGMapper to For the expression data of Egfr and its associated conduct such a search [16]. It is a software tool for genes we collected the expression data of Egfr axis automatically matching defined functional terms to from two sets of data, the male and the female sets. genes from a defined genome region by combining We used the Actin beta as controls for the Egfr. When the mapping information from the Ensembl database multiple probes were presented for the gene, the and gene function information from the OMIM and probe with the highest expression level was chosen PubMed databases. for the analysis while the others were used as reference. Additional Statistical Analysis The top 50 associated genes in each dataset on The association of the expression levels of Egfr the basis of Pearson correlation were used for plotting between female and male mouse populations Network Graphs in GeneNetwork. T–test was We conducted the association analysis using the

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 673 performed to test the levels of gene expression levels of different strains. We also conducted the same between the male and female mice. P value of 0.05 statistical analyses on the levels of Egfr expression was regarded as the largest value of indication of between the male and female mice. For Egfr, the P significant difference between the two populations. value for T test between the female and male is 1.07488E-17. The R value from the correlation analysis Results is 0.31. Therefore, there seems a difference in the expression levels between the male and female mice. Expression levels of Egfr between female and Thus, there is no difference in the expression level of male in the Liver Actin β between the female and male populations We conducted statistical analysis using data of while there is a significant sex difference in the Egfr gene expression of livers from a total of 39 strains, expression levels. Therefore, further analyses were which have both the data from Actin β (beta) and Egfr conducted as shown in the next several sections. in both sexes. There is one probe for Actin β on the Agilent G4121A chip. Figure 1A and 1B showed the Gene network of Egfr between female and expression levels of Actin β in female and male mice male in moue liver of different strains. For Actin B, the P value for T test Using the data on the expression level of probe between female and male is 0.88. The R value from the named A_51_P392242, the probe for Egfr and whole correlation analysis is 0.60. These data indicated that genome expression profiles in the livers of female RI there is a high correlation in the expression levels of strains, we identified the top 50 probes of genes with Actin β between male and female mice. Similarly, we their expression levels mostly correlated to that of Egfr obtained one probe for Egfr on the UNC Agilent (Supplementary Table S1). G4121A chip. Figure 2A and 2B showed the expression levels of Egfr in the female and male mice

Figure 1. Expression levels of Actin β in the liver between the female and male mice in the BXD strains. Numbers on Y-axis are for the relative levels of the expression of Figure 2. Expression levels of Egfr in the liver between the female and male mice in Actin β. The information of strains and names of standard inbred strains are listed the BXD strains. Numbers on Y-axis are for the relative levels of the expression of under the X-axis. A. The expression levels of Actin β in the liver of the female mice. Egfr. The information of strains and names of standard inbred strains are listed under B. The expression levels of Actin β in the liver of the male mice. the X-axis. A. The expression levels of Egfr in the liver of the female mice. B. The expression levels of Egfr in the liver of the male mice.

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 674

As shown in Figure 3A, Egfr is positively and Tram1 and Sec61b then have strong connections to negatively correlated to several genes in the female multiple genes. Egfr is strongly negatively correlated mice. The expression of mouse Egfr is strongly to dedicator of cytosolic acyl-CoA thioesterase 1 positively correlated to translocating (Cte1). Cte1 has negative impact on several other chain-associating membrane 1 (Tram1), genes. RIKEN cDNA 1190006C12 gene (for Sec61b); Both

Figure 3. Gene network of top 50 genes that are closely associated to Egfr in the male mouse liver and their network in the female liver. The 50 nodes in the graph below show the selected traits. Only nodes with the edges are displayed. The graph's canvas is 40.0 by 40.0 cm; the node labels are drawn with a 18.0 point font, and the edge labels are drawn with a 18.0 point font. A. Gene network in the female liver. The 684 edges between the nodes, filtered from the 1225 total edges and drawn as curves, show Pearson correlation coefficients greater than 0.5 or less than -0.5. B. Gen network in the liver of the male mice. The 366 edges between the nodes, filtered from the 1225 total edges and drawn as curves, show Pearson correlation coefficients greater than 0.5 or less than -0.5.

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 675

We then constructed a gene network in the liver 5 or 6 from the data of male liver. The eQTL for male of the male mice using the same probes from female mice are on and 8, but all suggestive. gene expression profiles of liver. As shown in Figure Thus, the level of Egfr expression is differently 3B, Egfr has even a stronger connection to C8b in regulated between male and female mice. female than that in male mice. It has no significant influence on the expression levels of Tram1, Cte1, and Sec61b in female. In addition, in male mice, Egfr has strong positive influence on the expression levels of E2F transcription factor 4 (E2f4); Forkhead box B1 (Foxb1); heterogeneous nuclear ribonucleoprotein K (Hnrnpk); IK cytokine (Ik); RAN binding protein 1 (Ranbp1); RIKEN cDNA 2700085A14 gene (Sas); CD2 antigen family, member 10 (Slamf9); small acidic protein (Smacp); Trk-fused gene (Tfg); and zinc finger protein 339 (Zfp339). In female there is no strong association between Egfr and these genes. The expression levels of key genes correlate with Egfr differently between female and male in the moue liver Based on these differences, we constructed second level of gene network between male and female mice. Figure 4A shows the positive and negative connections of these 15 genes to Egfr. The expression level of Egfr is connected to the expression level of each of these genes in the male liver. In female, as shown in Figure 4B, the expression of Egfr is not correlated with the expression level of any except the C8b. We further examined the correlation between the expression level of Egfr and several genes in the male and female mice. Specifically, in the male, the expression levels of Egfr is strongly positively correlated to the expression levels of E2f4 (Figure 5A) and Hnrnpk (Figure 5B) among the RI strains, while in the female there is no such a strong connection. In the Figure 4. Genes positively and negatively connected to the expression of Egfr in the male, the expression level of Egfr is strongly mouse liver. The 16 nodes in the graph below show the selected traits. All the nodes negatively connected to the expression of Cte1 (Figure are displayed. The graph's canvas is 40.0 by 40.0 cm; the node labels are drawn with a 16.0 point font, and the edge labels are drawn with a 16.0 point font. A. Genes 5C) and inhibitor of kappaB kinase gamma (Ikbkg) positively and negatively connected to the expression of Egfr in the male liver. The 69 (Figure 5D) while in the female there is no such a edges between the nodes, filtered from the 120 total edges and drawn as curves, show Pearson correlation coefficients greater than 0.5 or less than -0.5. B. Gene connection. These data confirm a significant network based on the expression level in the female liver for genes that are positively and negatively connected to the expression of Egfr. The 35 edges between the nodes, difference in the Egfr pathways between male and filtered from the 120 total edges and drawn as curves, show Pearson correlation female mice. coefficients greater than 0.5 or less than -0.5. Gene regulation of the expression of Egfr in the female and male liver Figure 6B shows the peak region of the eQTL on We finally examined the genetic factors that chromosome 5. The region between 141.00 and 142.50 regulate the expression of Egfr in the livers of female contains 12 genetic elements, including 10 known and male mice. With 5000 permutation test, we genes and two probes for other genetic components detected eQTLs for the regulation of Egfr from female (Table 1). In order to explore the possibility of any of liver on chromosome 5 and 6 (Figure 6A). The eQTL these 12 genes as the causal gene, we investigated two on chromosome 5 is significant. The significant aspects of these genes. We first examined the threshold is 18.48, and the suggestive threshold is correlations between the expression levels of these 10.87. The eQTL on chromosome 5 reached the level of genes and that of Egfr. Gene network indicated that 19. We did not detect an eQTL on either chromosome there is no connections between the expression level in any of these 12 genes and that of Egfr (Figure 6C). If

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 676 we concentrated on the very top of peak region of the eQTL on chromosome 5, which is located between 141.5 and 142.5 Mbp (Figure 5B), we obtained only 3 genetic elements as candidate genes. These candidates are sidekick homolog 1 (Sdk1), RIKEN cDNA 3200001G23 gene (3200001G23Rik), and KDEL (Lys-Asp-Glu-Leu) endoplasmic reticulum protein retention receptor 2 (Kdelr2). We next examined the polymorphisms of these three candidate genes between strains of B6 and D2 based on the Jackson laboratory database of Mouse Genome Informatics (http://www.informatics. jax.org/). We searched all the polymorphisms within each gene and 2 kb upstream and downstream of the gene. There are 23 polymorphic sites in Kdelr2 but none of them are different between D2 and B6. There are three polymorphisms in 3200001G23Rik (transcript for unknown gene) between D2 and B6. They are all located in the intron regions. There are 309 polymorphisms in Sdk1, all of which are SNPs. Most of these SNPs are within the intron while 4 of these are in the coding region. Three of these are synonymous while one is nonsynonymous polymorphism. The ID of the one nonsynonymous polymorphism is rs29564987. It is C/T missense polymorphism, where D2 codes for a “C” and B6 codes for a “T”. Accordingly, the Sdk1 is the most favorite candidate gene for the regulation of expression of Egfr in the liver of the mouse.

Figure 5. Correlation between the expression levels of Egfr and four key genes in the male mice. A. Egfr is strongly positively correlated to the expression of E2f4. B. Egfr is strongly positively correlated to the expression of Hnrnpk. C. Egfr is strongly negatively correlated to the expression of Cte1. D. Egfr is strongly negatively correlated to the expression of Ikbkg.

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 677

Discussion the other hand, our eQTL analysis did not map the loci for Egfr onto the position where the estrogen Our data showed that there are significant locates. Future studies may reveal whether this is differences in the molecular pathways of Egfr in the because that in the liver the estrogen does not play a liver between female and male mice. The data agree major role for the sex differential expression. The with the fact that the Egfr is regulated by estrogen [3, significant difference in the molecular pathways of 17]. Our data further indicated that the sex differences Egfr between the female and male mice in these exist not only in the molecular pathways in which Egfr strains has laid a foundation for this important study. regulates and interacts with other genes but also in the regulation mechanisms of Egfr by other genes. On

Figure 6. eQTL map of Egfr based on the whole genome expression profiles of the liver of the female population. A. EQTL map of whole genome showing the loci on chromosome 5, 6, and 18. B. The peak region of eQTL on chromosome 5. C. Gene network of 12 candidate genes of eQTL on chromosome 5 and Egfr derived from the liver of the female mice. The 13 nodes in the graph below show the selected traits. All the nodes are displayed. The 15 edges between the nodes, filtered from the 78 total edges and drawn as curves, show Pearson correlation coefficients greater than 0.3 or less than -0.3. The graph's canvas is 40.0 by 40.0 cm; the node labels are drawn with a 16.0 point font, and the edge labels are drawn with a 16.0 point font.

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 678

Table 1. Information of candidate genes for eQTL of expression level of Egfr in male mice.

Symbol Mb Start Length (Kb) SNP SNP Density Avg.Expr Gene Description (count) Chst12 140.9816 19.63 2 0.101885 1.287 Carbohydrate sulfotransferase 12 Grifin 141.0391 1.879 0 0 0.681 Galectin-related inter-fiber proteinprovided Lfng 141.0833 8.205 0 0 0.077 LFNG O-fucosylpeptide 3-beta-N-acetylglucosaminyltransferase Ttyh3 141.0966 28.385 1 0.03523 - Tweety homolog 3 (Drosophi... Iqce 141.1395 38.874 7 0.180069 -- IQ motif containing E AA881470 141.181 14.31 2 0.139762 -- EST AA881470 Amz1 141.2001 29.186 4 0.137052 -- Archaelysin family metallopeptidase 1 Gna12 141.2359 70.488 21 0.297923 0.691 Guanine nucleotide binding protein, alpha 12 Card11 141.349 127.598 24 0.188091 -- Caspase recruitment domain family, member 11 Sdk1 141.7175 972.258 605 0.622263 -- Sidekick homolog 1 (chicken)... 3200001G23Rik 142.0698 1.197 0 0 -- RIKEN cDNA 3200001G23 gene Kdelr2 142.5 0.05 0 0 1.070 KDEL (Lys-Asp-Glu-Leu) endoplasmic reticulum protein retention receptor 2

Most of the sex differentially regulated genes in expression. Although the differential expression this study have not been investigated for liver cancer. levels of Egfr between the female and the male mice However, a notable feature of these genes is their apparently is not regulated by estrogen, the difference critical roles in the pathways of gender specific in the gene network of Egfr between the female and cancers. For example, E2f4 have been reported to the male are mainly caused by genes in the estrogen regulate liver cancer cell survival and risk for breast pathway. For example, E2f4 and Hnrnpk are well cancer [18-20]; Hnrnpk has been linked to the known by their interactions with estrogen. In neuroendocrine differentiation signaling in prostate addition, Lasarte et al. reported that estradiol cancer cells [21] as well as to the development of downregulates NF-κb translocation by Ikbkg breast cancer [22]. In brain cancer the expression of transcriptional repression in dendritic cells [27]. Hnrnpk has been linked to the Egfr [23]. The increased Furthermore, the relation between estrogen and Sdk1 transcription level of Ranbp1 has been linked to the has not been studied. There is a possibility that Sdk1 decreases taxol sensitivity in RKO colon carcinoma has a role in the estrogen pathway. Nevertheless, cells [24]. Remarkably, Ikbkg has been linked to several further study is necessary before any conclusion is cancers and its cytokine network has been the target reached. for drug development for breast cancer [25]. In humans, although there is no intensive Furthermore, Ikbkg and Egfr have been reported to investigation on the gender difference of anti EGFR upregulate PKM2 expression and promote therapy for the liver cancer, there are few reports tumorigenesis [26]. Understanding the gender showing the potential gender difference. For example, differences of their functions in the cancer genetics there is apparently a gender difference in hazard ratio and responses to drug treatment will greatly enhance (HR) in patients in a phase III, randomized, the future drug development and therapeutic double-blind, placebo-controlled trial of sorafenib application. plus erlotinib [9] in patients with advanced Remarkably, the regulation of the Egfr hepatocellular carcinoma [28]. While the HR of overall expression level between female and male mice survival (OS) in the female patients is 1.135 showed much difference. In the male, there was no (0.708-1.821), the male patients is 0.926 (0.762-1.126). significant loci for the regulation of the expression Brivanib is an ATP-competitive inhibitor against level of Egfr, while in the female, there were at least VEGFR2, moderate potency against VEGFR-1 and two loci being identified. According to our data, Sdk1 FGFR-1. In a study using brivanib as adjuvant therapy is the most favorite candidate for the regulation of to transarterial chemoembolization in the patients expression level of Egfr in the female mice. Although with hepatocellular carcinoma, Kaplan-Meier's curves some researches on the molecular pathways of Sdk1 for OS in the female patients is 1.71 (0.78-3.74), but in have been done, it has not been directly linked to the Male patients is 0.79 (0.56-1.11) [29]. The evidences cancer. It is possible that it may play a secondary role and the data from animal studies again emphasize the in the cancer development by regulation of expression importance of gender difference not only in disease level of Egfr differentially between female and male. incidence but also in response to drug treatment and On the other hand, the data may indicate that there is overall survival in the liver cancer patients. a complicated regulation of Egfr expression between One limitation of this study is that our data is at the female and the male, in some cases, there may be one time point. Future studies at different age stages different in binding sites in the regulation of Egfr may lead to different results and comprehensive

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 679 understanding of the sex difference in the molecular 2. Yarden Y. The Egfr family and its ligands in human cancer. signalling mechanisms and therapeutic opportunities. Eur J Cancer. 2001;37 (Suppl pathways of Egfr axis. In addition, our study is based 4):S3-8. 3. Filardo EJ. Epidermal growth factor receptor (Egfr ) transactivation by on the RI strains that were derived from two mouse estrogen via the G-protein-coupled receptor, GPR30: a novel signaling strains. Mouse strain specific pathways have been pathway with potential significance for breast cancer. J Steroid Biochem Mol Biol. 2002;80(2):231-8. known [30-31]. Nevertheless, the two parental strains, 4. Lanaya H, Natarajan A, Komposch K, Li L, Amberg N, Chen L, Wculek SK, C57BL/6J and DBA/2J, are the strains that have been Hammer M, Zenz R, Peck-Radosavljevic M, Sieghart W, Trauner M, Wang H, Sibilia M. Egfr has a tumour-promoting role in liver macrophages during widely used in the biomedical research. We feel hepatocellular carcinoma formation. Nat Cell Biol. 2014;16(10):972-81, 1-7. doi: confident on the reliability and accuracy of the data. 10.1038/ncb3031. Epub 2014 Aug 31. 5. Malamitsi-Puchner A, Tziotis J, Tsonou A, Protonotariou E, Sarandakou A, The raw data in this study were from multiple Creatsas G. Changes in serum levels of vascular endothelial growth factor in homozygous RI strains [12-13]. Gene expression males and females throughout life. J Soc Gynecol Investig. 2000; 7:309-312. 6. Baserga M, Bares AL, Hale MA, Callaway CW, McKnight RA, Lane PH, Lane profiles from each strain were generated from RH. Uteroplacental insufficiency affects kidney VEGF expression in a model of multiple mice. Mice of both sexes were from the same IUGR with compensatory glomerular hypertrophy and hypertension. Early Hum Dev. 2009; 85:361-367. strain with the same genomic background except that 7. Brandon-Warner E, Walling TL, Schrum LW, et al. Chronic ethanol feeding they differed by sex. All the animals were kept in the accelerates hepatocellular carcinoma progression in a sex-dependent manner in a mouse model of hepatocarcinogenesis. Alcohol Clin Exp Res same animal facility with the same environment. The 2012;36(4):641-653. 8. Lee JK, Abe K, Bridges AS, Patel NJ, et al. Sex-dependent disposition of none-difference of the expression levels of Actin B acetaminophen sulfate and glucuronide in the in situ perfused mouse liver. between female and male mice in these strains Drug Metab Dispos. 2009;37(9): 1916-1921. 9. Hsu CH, Kang YK, Yang TS, Shun CT, Shao YY, Su WC, Sandoval-Tan J, demonstrated the reliability of the data. Chiou TJ, Jin K, Hsu C, Cheng AL. Bevacizumab with erlotinib as first-line We are aware that notable differences between therapy in Asian patients with advanced hepatocellular carcinoma: a multicenter phase II study. Oncology. 2013;85(1):44-52 the mouse models and humans have been found [32]. 10. Bekaii-Saab T, Markowitz J, Prescott N, Sadee W, Heerema N, Wei L, Dai Z, The differentially expressed Egfr between the female Papp A, Campbell A, Culler K, Balint C, O'Neil B, Lee RM, Zalupski M, Dancey J, Chen H, Grever M, Eng C, Villalona-Calero M. A multi-institutional and male mice may not be differentially expressed phase II study of the efficacy and tolerability of lapatinib in patients with between men and women in humans. However, our advanced hepatocellular carcinomas. Clin Cancer Res. 2009;15(18):5895-5901. 11. Jiao Y, Chen H, Yan J, Wang L, Huang Y, Liu X, Williams RW, Lu L, Wang Y, data will serve as a reminder that it is important to Gu W. Genome-wide gene expression profiles in antioxidant pathways and keep in mind on the gender difference in the drug their potential sex differences and connections to vitamin C in mice. Int J Mol Sci. 2013;14(5):10042-10062. development targeting EGFR and its family members. 12. Gatti DM, Zhao N, Chesler EJ, Bradford BU, Shabalin AA, Yordanova R, Lu L, Rusyn I. Sex-specific gene expression in the BXD mouse liver. Physiol Genomics. 2010;42(3):456-468. Supplementary materials 13. Gatti D, Maki A, Chesler EJ, Kirova R, Kosyk O, Lu L, Manly KF, Williams RW, Perkins A, Langston MA, Threadgill DW, Rusyn I. Genome-level analysis Supplementary Tables. of genetic regulation of liver gene expression networks. Hepatology. http://www.jcancer.org/v07p0671s1.xlsx 2007;46(2):548-57. 14. Wang L, Liu H, Jiao Y, Wang E, Clark SH, Postlethwaite AE, Gu W, Chen H. Differences between Mice and Humans in Regulation and the Molecular Acknowledgements Network of Collagen, Type III, Alpha-1 at the Gene Expression Level: Obstacles that Translational Research Must Overcome. Int J Mol Sci. This work was partially supported by the Center 2015;16(7):15031-56. 15. Wang L, Jiao Y, Sun S, Jarrett HW, Sun D, Gu W. Gene network of a of Genomics and Bioinformatics and Center of phosphoglycerate mutase in muscle wasting in mice. Cell Biol Int. Connective Tissue Research at the University of 2015;39(6):666-77. 16. Xiong Q, Qiu Y, Gu W. PGMapper: a web-based tool linking phenotype to Tennessee Health Science Center; The collaborative genes. Bioinformatics. 2008;24(7):1011-3. project between the first hospital of Qiqihare city and 17. Aizen J, Thomas P. Role of Pgrmc1 in estrogen maintenance of meiotic arrest in zebrafish oocytes through Gper/Egfr. J Endocrinol. 2015;225(1):59-68. the University of Tennessee health Science Center. 18. Rodríguez JL, Sandoval J, Serviddio G, Sastre J, Morante M, Perrelli MG, Martínez-Chantar ML, Viña J, Viña JR, Mato JM, Avila MA, Franco L, Author contributions López-Rodas G, Torres L. Id2 leaves the chromatin of the E2F4-p130-controlled c-myc promoter during hepatocyte priming for liver WG and HC designed the experiments. LW, JX regeneration. Biochem J. 2006;398(3):431-7. 19. Ferrer-Vicens I, Riffo-Campos ÁL, Zaragozá R, García C, López-Rodas G, Viña and WG analyzed the data, constructed the gene JR, Torres L, García-Trevijano ER. In vivo genome-wide binding of Id2 to E2F4 networks, and mapped the eQTLs. HC and WG target genes as part of a reversible program in mice liver. Cell Mol Life Sci. 2014;71(18):3583-97. conceived and managed the project. LW, JX and WG 20. González-Hernández A, Henríquez-Hernández LA, Cabrera de León A, contributed to the manuscript preparation. All Rodríguez-Pérez Mdel C, Murias-Rosales A, Domínguez-Coello S, Brito-Díaz B, Almeida-González D, Aguirre-Jaime A, Díaz-Chico BN. Microsatellite authors contribute to the manuscript finalization. polymorphisms in the EGFR, NOTCH4 and E2F4 genes and their association with breast cancer risk. Int J Biol Markers. 2012;27(3):e219-26. doi: 10.5301/JBM.2012.9583. Conflict of interest 21. Ciarlo M, Benelli R, Barbieri O, Minghelli S, Barboro P, Balbi C, Ferrari N. The authors declare that there are no conflicts of Regulation of neuroendocrine differentiation by AKT/hnRNPK/AR/β-catenin signaling in prostate cancer cells. Int J Cancer. interest. 2012;131(3):582-90 22. Hamrita B, Chahed K, Kabbage M, Guillier CL, Trimeche M, Chaïeb A, Chouchane L. Identification of tumor antigens that elicit a humoral immune References response in breast cancer patients' sera by serological proteome analysis 1. Voldborg BR, Damstrup L, Spang-Thomsen M, Poulsen HS. Epidermal growth (SERPA). Clin Chim Acta. 2008;393(2):95-102. factor receptor (Egfr) and Egfr mutations, function and possible role in clinical 23. Babic I, Anderson ES, Tanaka K, Guo D, Masui K, Li B, Zhu S, Gu Y, Villa GR, trials. Ann Oncol. 1997;8(12):1197-206. Akhavan D, Nathanson D, Gini B, Mareninov S, Li R, Camacho CE, Kurdistani SK, Eskin A, Nelson SF, Yong WH, Cavenee WK, Cloughesy TF, Christofk HR, Black DL, Mischel PS. EGFR mutation-induced alternative splicing of Max

http://www.jcancer.org Journal of Cancer 2016, Vol. 7 680

contributes to growth of glycolytic tumors in brain cancer. Cell Metab.2013;17(6):1000-8. 24. Amato R, Scumaci D, D'Antona L, Iuliano R, Menniti M, Di Sanzo M, Faniello MC, Colao E, Malatesta P, Zingone A, Agosti V, Costanzo FS, Mileo AM, Paggi MG, Lang F, Cuda G, Lavia P, Perrotti N. Sgk1 enhances RANBP1 transcript levels and decreases taxol sensitivity in RKO colon carcinoma cells. Oncogene. 2013;32(38):4572-8. 25. Barbie TU, Alexe G, Aref AR, Li S, Zhu Z, Zhang X, Imamura Y, Thai TC, Huang Y, Bowden M, Herndon J, Cohoon TJ, Fleming T, Tamayo P, Mesirov JP, Ogino S, Wong KK, Ellis MJ, Hahn WC, Barbie DA, Gillanders WE. Targeting an IKBKE cytokine network impairs triple-negative breast cancer growth. J Clin Invest. 2014;124(12):5411-23. 26. Yang W, Xia Y, Cao Y, Zheng Y, Bu W, Zhang L, You MJ, Koh MY, Cote G, Aldape K, Li Y, Verma IM, Chiao PJ, Lu Z. EGFR-induced and PKCε monoubiquitylation-dependent NF-κB activation upregulates PKM2 expression and promotes tumorigenesis. Mol Cell. 2012;48(5):771-84. 27. Lasarte S1, Elsner D, Sanchez-Elsner T, Fernandez-Pineda A, López-Fernández LA, Corbí AL, Muñoz-Fernandez MA, Relloso M. Estradiol downregulates NF-κb translocation by Ikbkg transcriptional repression in dendritic cells. Genes Immun. 2013;14(7):462-9. doi: 10.1038/gene.2013.35. Epub 2013 Jul 25. 28. Zhu AX, Kudo M, Assenat E, et al. Effect of everolimus on survival in advanced hepatocellular carcinoma after failure of sorafenib, the EVOLVE-1 randomized clinical trial. JAMA 2014; 312(1):57-67. 29. Kudo M, Han G, Finn RS, et al. Brivanib as adjuvant therapy to transarterial chemoembolization in patients with hepatocellular carcinoma, A randomized phase III trial. Hepatology 2014; 60(5):1697-1707 . 30. Mo Y, Chen J, Schlueter C.F, Hoyle G.W. Differential susceptibility of inbred mouse strains to chlorine-induced airway fibrosis. Am J Physiol Lung Cell Mol Physiol. 2013; 304: L92–L102. 31. Haston, C.K.; Cory, S.; Lafontaine, L.; Dorion, G.; Hallett, M.T. Strain-dependent pulmonary gene expression profiles of a cystic fibrosis mouse model. Physiol Genomics 2006; 25:336–345. 32. Wang L, Liu H, Jiao Y, Wang E, Clark SH, Postlethwaite AE, Gu W, Chen H. Differences between Mice and Humans in Regulation and the Molecular Network of Collagen, Type III, Alpha-1 at the Gene Expression Level: Obstacles that Translational Research Must Overcome. Int J Mol Sci. 2015;16(7):15031-56.

http://www.jcancer.org